In general, computers and people come into contact with each other through a keyboard/display unit. Thus, whether it is easy to handle the keyboard/display unit will greatly affect the efficiency of the work.
Recently, there have appeared advanced terminal devices with which processing of Japanese characters including Kanji characters (Chinese characters) is possible. With these terminals, the computer input/output of data consisting of Japanese characters is much easier. There are still, however, some weak points in processing data of Japanese Kanji characters. First, there are too many characters to be entered efficiently by key operation. These characters cannot be entered as easily as Latin or English alphabetic characters are entered.
Second, a worldwide common code system consisting of the same number of bits is not in use, since Kanji characters require a 2-byte (16 bit) code system, while 7-bit code systems (e.g. ASCII code) have been adopted for computers used in almost all nations in the world with, for example, the characters of the Latin or English alphabet. This prevents smooth communication of data processing information between Japan and foreign countries.
An 8-bit code system JIS C 6220, created on the basis of the ASCII code, is available. It is, however, impossible to encode a 7-bit software code for JIS C 6220 without modification. Furthermore, since JIS C 6220 is based on Katakana characters, documents output through this code are very difficult to read, and many readers of those Katakana documents complain about this difficulty.
Texts consisting of a Latin or English alphabet, in combination with Hiragana and Katakana characters, present special problems. The readability of such documents has been studied. In a text, for example, consisting mainly of Hiragana characters, it may be read as if it were normal Japanese writing with spaces inserted between words. In conventional systems, however, a code of many bits is required if the three types of character sets are used, even though the number of bits used for writing would be less than the case in which Kanji characters are used.
There are still some other systems in which many characters are represented by codes consisting of fewer bits, and shift codes are set in advance in the control code. For example, in the ASCII code, the hexadecimal numbers "OE" and "OF" correspond to SO (shift out) and SI (shift in), respectively, and the succeeding character sets are selected by these control codes (SO/SI), for example, upper/lower cases.
In a text in which Latin or English characters, Hiragana characters and Katakana characters are mixed and frequently change, the following problems occur:
(1) The shift system is merely used to change character sets. Thus, it is impossible to judge what character set is used to represent the word by simply checking the word code system after it has switched over. In some cases, the determination may not be made before thousands of characters have been traced back to check the SI or SO.
(2) If three types of characters are used in this system, three change-over codes must be set. In this case, problem (1) above may be worsened. It becomes almost impossible to use the code.
(3) In programs created mainly by the ASCII code, SI and SO are often used for purposes other than selection of graphic symbols, etc. Thus SI and SO, if used for changing between Hiragana and Katakana characters, cannot retain compatibility with conventional programs.
In actuality, they are rarely used because of these problems.
A recently developed shift system used in data processing of Japanese characters, including Kanji characters, has come into use. It is an extension of the above shift code system and has the same basic concept as the above SI and SO system, having the same problems as (1), (2) and (3) listed above.